Skip to main content
def NumEncoderWorker.fit(self, inputs, fx_choice=REG.AIG, **params) [source]
Analyze and learn numerical bin encoding. Executes the non-greedy bin search, min-bin-share merging, and predictive gates for each numerical feature. Results are stored as encoding maps ready for transform.

​
Parameters

inputs : NumEncoderInput
Input configuration including training set and
numerical feature lists.
 fx_choice : str
The numerical encoding strategy. Defaults to REG.AIG.
 **params : dict
Runtime parameters passed to the internal fitter
(max_num_bins, gini_cutoff, num_alpha, min_bin_share).

​
Returns

NumEncoder
The fitted encoder instance.